Uploaded February 2022 | Updated September 2026, 1 week ago

---

Wei Wei, a Developer Advocate for TensorFlow, discusses continuous and discrete control, and walks through how to use the Minitaur environment to train a 4-legged robot to walk with the Actor-Learner API.

Resources:
SAC minitaur with the Actor-Learner API Tutorial → https://goo.gle/3oAa9FF

TF Agents SAC example → https://goo.gle/35OzF3u

TF Agents distributed training → https://goo.gle/3rzYj06

TensorFlow actor-critic tutorial → https://goo.gle/3HFckiN


Chapters:
00:00 Introduction
00:27 SAC Agent
02:01 The minitaur environment
03:20 Minitaur training tutorial
07:05 TF-Agents distributed collection
07:49 Distributed architecture
08:30 Summary and resources


Watch more Reinforcement learning with TensorFlow Agents episodes → https://goo.gle/reinforcement-learning
Subscribe to TensorFlow → https://goo.gle/TensorFlow
Ask your questions on the TF Forum → https://goo.gle/discuss_tensorflow



#TensorFlow #MachineLearning #ML


product: TensorFlow - General; fullname: Wei Wei;